Given the growing size of patent databases and portfolios, scholars and practitioners alike need metrics to help them in weighting patent counts or in ranking patents to focus on the most important ones. The main objective of this paper is to help them in this task with a practical and replicable approach. The discriminating criterion chosen is the potential market for the patented invention and the approach uses five different patent features (forward citations, grant decisions, families, renewals, and oppositions) that (1) have been found positively correlated with the value of patents in the literature, (2) can be extracted from patent databases, and (3) could inform on the existence of a potential market for the invention. The paper therefore discusses the main methodological issues in measuring and interpreting these metrics over a large international dataset and proposes one approach to extract from the different measures a consistent score that can be used to weight or rank patents.
AbstractGiven the growing size of patent databases and portfolios, scholars and practitioners alike need metrics to help them in weighting patent counts or in ranking patents to focus on the most important ones. The main objective of this paper is to help them in this task with a practical and replicable approach. The discriminating criterion chosen is the potential market for the patented invention and the approach uses five different patent features (forward citations, grant decisions, families, renewals, and oppositions) that (1) have been found positively correlated with the value of patents in the literature, (2) can be extracted from patent databases, and (3) could inform on the existence of a potential market for the invention. The paper therefore discusses the main methodological issues in measuring and interpreting these metrics over a large international dataset and proposes one approach to extract from the different measures a consistent score that can be used to weight or rank patents.